Rewrite sessionNameGenerator.ts to consume a StreamFn-shaped dependency
(sourced from PiAgentSession.agent.streamFn, wired in leg B) instead of
ModelRegistry plus a dynamic @earendil-works/pi-ai/compat import.
streamFn resolves auth/headers/retry internally, so the explicit
modelRegistry.getApiKeyAndHeaders(model) call and apiKey/headers stream
options are no longer needed.
Deletes now-dead compat-loading machinery: getPiAiProviderRegistryModule,
loadPiAiProviderRegistryModule, importOptionalPiAiModule,
isModuleUnavailableError, hasGetApiProvider, PI_AI_COMPAT_MODULE,
ModuleImporter, SessionNameApiProvider, PiAiProviderRegistryModule, and
the module-level provider registry cache.
Updates the maybeGenerateSessionName call site in piSessionService.ts to
pass session.agent.streamFn instead of this.modelRegistry, and updates
both test files: sessionNameGenerator.test.ts gains coverage for the new
streamFn-driven generateShortSessionName signature (success and error
paths), and piSessionService.test.ts gains an end-to-end test proving a
first prompt generates a session name through the wired agent.streamFn
fake.
sessionNameGenerator.ts no longer imports @earendil-works/pi-ai/compat
or references getApiProvider anywhere.
Add a narrow `agent: { streamFn: StreamFn }` field to the PiAgentSession
interface, sourced structurally from the real AgentSession.agent (Agent,
from the stable @earendil-works/pi-agent-core package). This is the
resolved-auth/headers/retry "call this model" function pi's own
compaction/branch-summarization code uses internally, and will let
sessionNameGenerator.ts drop its dependency on the deprecated pi-ai
/compat provider registry in a follow-up leg.
No adapter object was needed: AgentSession's real .agent field already
structurally satisfies the new narrow shape, so defaultCreateAgentRuntime
needed no changes.
Adds a unit test proving the field is wired end-to-end via the injected
runtime creator. sessionNameGenerator.ts itself is untouched (leg C).
Split subsession inspection into two tools: check_subsession gives a quick
glance (status + latest assistant output), while read_subsession reads through
a child's transcript with role/content filters, full-content substring search
(including tool-call args), optional per-value maxChars truncation that flags
clipped parts, includeToolArgs, and pagination.
Filtering and search run on full untruncated content; truncation is an explicit,
caller-owned final projection (no default) so a narrow read never silently hides
a match. Empty page-windows are distinguished from zero-match results.
Add spawn_subsession / list_subsessions / read_subsession tools that let an
agent start child sessions it stays attached to: the child records its parent
in the session tree, the parent is notified (as a system-authored custom
message that wakes an idle parent and queues behind in-flight work) when the
child stops working, and the parent can inspect children's status and result.
Gated behind a beta flag, off by default, mirroring spawnSessions: enable via
PI_WEB_SUBSESSIONS, the subsessions config key, or the Settings toggle. Also
requires spawnSessions.
Also fix the release skill so the version step resyncs package-lock.json
(npm install --package-lock-only) and the commit step refuses a release where
package.json and package-lock.json versions disagree.
Attachments can be pasted, dropped, or uploaded, so 'paste' was too
narrow. Rename the default folder and filename prefix accordingly and
update the changeset reference.
A slash command sent as the first (or any idle) message showed the raw
command text twice until reload: once from the client's optimistic insert
and once from the server's message.append echo, neither of which converges
with the agent's canonical expanded message (e.g. a /skill:* block).
Make commands obey the same source-of-truth contract as prompts:
- client no longer inserts the raw command text optimistically; it shows
the existing per-session sending indicator instead
- forwarded runtime/skill commands return a bare done result rather than a
synthetic "Accepted ..." line
- server suppresses the raw message.append echo for command-forwarded
prompts (threaded through the compaction queue too)
Result: pre-reload state matches reload, with no transient duplicate.
Add a project-scoped spawn_session tool so agents can dispatch new,
independent sessions (ralph loops, long-plan chaining). Spawned sessions
are constrained to a workspace/worktree of the same registered project,
appear in the session list immediately via a new session.created event,
and the capability is on by default with a Settings -> Session daemon
toggle (spawnSessions / PI_WEB_SPAWN_SESSIONS).
Note: adds a session daemon code path, so pi-web-sessiond.service must be
restarted manually for the server side to take effect.
Builds on marcus's plugin-api-completeness work. Narrows the new plugin
surface to capabilities that expose real, otherwise-unreachable pi-web
functionality, and drops invented/duplicative surfaces:
Kept:
- files.writeFile / deleteFile / moveFile (genuine workspace mutation,
federated, path-safe)
- prompt.insertText / getText / getSelection (editor state access)
Dropped:
- attachments.* (insertFileReference/getAttachedFiles/removeFileReference):
getAttachedFiles invented a structured-attachment notion pi-web does not
have and duplicated prompt.getText() + a regex with a false email-safety
claim; insert/removeFileReference were thin sugar over readFile +
insertText that plugins can compose themselves.
- prompt.onPaste / onKeyDown: an incomplete two-event hook system shaped
around a single use case, overlapping the editor's native image-paste
handling. Deferred until a real editor event/hook surface is designed.
- prompt.focus: redundant and buggier duplicate of the existing
focusPrompt() (silently no-ops when not on the chat view). Focus stays
as focusPrompt().
Security fix:
- deleteWorkspaceFile now resolves the parent via realpath + ensureInside
before lstat/unlink, closing a symlinked-parent-directory escape that
allowed deleting files outside the workspace (write/move already did
this). Final path component is still not resolved, so deleting a symlink
removes the link, not its target. Adds a regression test.
Docs and the registry test mock updated to match the trimmed surface.
